Instructions to run each script

Requirements for running

1. Python 3.8
2. Pandas library
3. Numpy library

Code to run each part

Part 2

To train the emission parameters, run the following: python3 part2.py [dataset] train
Where [dataset] can be EN, SG, or CN.

The emission parameters will be saved as a pickle file (params.pkl) in the respective dataset folder.

To predict with the trained parameters, run the following: python3 part2.py [dataset] predict
Where [dataset] can be EN, SG, or CN.

The output file is saved as dev.p2.out in the respective dataset folder.

Part 3

Part 3 requires the emission parameters from part 2 (which is saved as a pickle file after training).

To train the transition parameters, run the following: python3 part3.py [dataset] train
Where [dataset] can be EN, SG, or CN.

The transition parameters will be saved as a pickle file (y_params.pkl) in the respective dataset folder.

To predict with the trained parameters, run the following: python3 part3.py [dataset] predict
Where [dataset] can be EN, SG, or CN.

The output file is saved as dev.p3.out in the respective dataset folder.

Part 4

Part 4 requires the emission parameters from part 2 and the transition parameters from part 3 (which is saved as a pickle file after training).

To convert the parameters to their respective dictionaries, run the following: python3 part4.py EN train

The dictionary for emission parameters is saved as em_dic.p and the dictionary for transition parameters is saved as tr_dic.p.

To output the 3rd best sequence, run the following: python3 part4.py EN predict

The output file is saved as dev.p4.out in the EN folder.

Part 5

Part 5 requires the emission parameters from part 2 and the transition parameters from part 3 (which is saved as a pickle file after training).

To convert the parameters to their respective dictionaries, run the following (you can skip this step if you have done it in part 4): python3 part5.py EN train

The dictionary for emission parameters is saved as em_dic.p and the dictionary for transition parameters is saved as tr_dic.p.

To output the 3rd best sequence, run the following: python3 part5.py EN predict

The output file is saved as dev.p5.out in the EN folder.

To output the 3rd best sequence for the test dataset, run the following:
python3 part5.py EN test

The output file is saved as test.p5.out in the EN folder.